Thermal Comfort Prediction Research Articles

An indoor thermal comfort model for group thermal comfort prediction based on K-means++ algorithm

An indoor thermal comfort model for group thermal comfort prediction based on K-means++ algorithm

The Mechanism of Street Spatial Form on Thermal Comfort from Urban Morphology and Human-Centered Perspectives: A Study Based on Multi-Source Data

The influence of street spatial form on thermal comfort from urban morphology and human-centered perspectives has been underexplored. This study, utilizing multi-source data and focusing on urban central districts, establishes a refined index system for street spatial form and a thermal comfort prediction model based on extreme gradient boosting (XGBoost) and Shapley additive explanations (SHAP). The results reveal the following: (1) Thermal comfort levels display spatial heterogeneity, with areas of thermal discomfort concentrated in commercial zones and plaza spaces. (2) Compared to the human-centered perspective, urban morphology indicators correlate strongly with thermal comfort. (3) The key factors influencing thermal comfort, in descending order of importance, are distance from green and blue infrastructure (GBI), tree visibility factor (TVF), street aspect ratio (H/W), orientation, functional diversity indices, and sky view factor. All but the TVF negatively correlates with thermal comfort. (4) In local analyses, the primary factors affecting thermal comfort vary across streets with different heat-risk levels. In high heat-risk streets, thermal comfort is mainly influenced by distance from GBI, H/W, and orientation, whereas in low heat-risk streets, vegetation-related factors dominate. These findings provide a new methodological approach for optimizing urban thermal environments from both urban and human perspectives, offering theoretical insights for creating more comfortable cities.

Enhancing indoor thermal comfort prediction in tropical regions: A transfer learning strategy in West Bengal

Ensuring energy-efficient building management and sustainability necessitates precise thermal comfort prediction, a task particularly critical in tropical regions with significant energy-saving potential. This study confronts the complexities introduced by humid air conditions and escalating internal and external heat levels. The objective is to devise a sophisticated strategy for predicting indoor personal thermal comfort that enhances prediction accuracy. The proposed method extends beyond previous techniques by incorporating a comprehensive array of thermal comfort factors and their potential interdependencies, while also leveraging a transductive transfer learning strategy to circumvent data sparsity in the target domain. The experimental results indicate the presence of localized feature coupling within thermal comfort datasets, which the proposed CNN-ConvTrans model effectively processes, thereby enhancing predictive efficacy. The primary innovation of this work lies in the development of a novel transfer learning strategy that considers localized feature coupling, integrated with a deep learning model, effectively addressing issues such as class imbalance and data scarcity. This study furnishes an innovative and valuable framework for thermal comfort prediction across diverse environmental settings.

Machine learning-based prediction of indoor thermal comfort in traditional Chinese dwellings: A case study of Hankou Lifen

Traditional residential dwellings as architectural heritage holds substantial research value and significance for regional studies. This study investigates the indoor thermal comfort of traditional “Hankou Lifen” dwellings compared to new residences during summer through field measurements and questionnaires. The results demonstrate that the indoor thermal environment of “Lifen” houses is inferior, leading to lower thermal comfort for the residents. A linear regression model assesses summer residents' thermal comfort, considering psychological perspectives to calculate natural and acceptable temperature ranges. However, these models primarily rely solely on temperature and overlook other environmental and personal factors, due to their lack of accuracy in capturing complex non-linear relationships. To address these limitations, six machine learning models with high predictive performance are selected for analyses. Data preprocessing techniques, such as upsampling, and hyper-parameter optimization are employed to enhance prediction performance. The selection of algorithms is based on metrics such as accuracy, precision, recall, and F1 score, with further performance evaluation conducted using ROC curves and confusion matrices. Through evaluation, XGBoost and LightGBM show superior performance in predicting residential thermal comfort.Finally, SHAP(SHapley Additive exPlanations) values and PDP(Partial Dependence Plot) are used to interpret the models, revealing various factors influencing residents' thermal sensations in both new and old dwellings. These findings emphasize the significance of thermal comfort in the preservation and revitalization of historic dwellings and propose a novel framework for advancing the sustainable utilization of traditional dwellings.

Developing a new adaptive heat balance model to enhance thermal comfort predictions and reduce energy consumption

Accurate thermal comfort prediction is essential for enhancing both thermal comfort and energy efficiency in buildings. The heat balance and adaptive models, while widely used, have been often questioned in relation to their strengths and limitations. To overcome these challenges, a comprehensive understanding of both the physical parameters influencing heat exchange and occupants’ adaptive capacities is essential.This study introduces an analytical method to formulate a new adaptive heat balance model, the adPMV, which integrates the strengths of both models considering various parameters influencing thermal perception. The model integrates the PMV with an adaptive factor associated with the running mean outdoor temperature, in line with adaptive theory.Tested on 1377 samples from European university classrooms, the adPMV demonstrates enhanced accuracy (MAE=0.74, RMSE=1.01, MBE=-0.11) compared to PMV and other adaptive heat balance models. Validation on naturally ventilated university classrooms from ASHRAE’s databases further confirms promising results, showcasing reduced error indices (MAE=0.68, RMSE=0.81, MBE=-0.03). Notably, using adPMV setpoints not only improves thermal sensation prediction accuracy but also leads to a substantial reduction in heating demand, reaching up to 40 %.The adaptability of this model to different contexts, such as building types, climates, and HVAC system operations, presents it as a versatile tool for exploring adaptive principles.

Developing building-specific, occupant-centric thermal comfort models: A methodological approach

This study addresses the research problem of accurately predicting thermal comfort in buildings by developing occupant-centric, building-specific models. Traditional models like PMV and adaptive models often fall short in reflecting individual comfort preferences. This research aims to create more accurate and personalised thermal comfort models by considering human-related factors such as clothing, mood, and activities, along with environmental variables like temperature and humidity. The methodology involved collecting data through two self-reporting campaigns using mobile and smartwatch applications in a university building in The Hague. Regression and classification models were developed, achieving accuracy rates of 72 % and 89 % respectively, which surpass the performance of traditional models. The findings indicate that human factors significantly influence thermal comfort, with mood and clothing being particularly impactful. Seasonal variations were also accounted for, emphasising the need for periodic data collection to capture changes in occupant behaviour and environmental conditions. The key contribution of this research lies in its ability to enhance the accuracy of thermal comfort predictions, leading to more effective and user-centric HVAC system operations. This approach not only has the potential to improve the comfort of building occupants but also has implications for energy efficiency, as HVAC systems can be better tailored to actual comfort needs.

Analyzing sleep thermal comfort with an attention-based gated recurrent unit neural network

To provide a comfortable sleep environment, the air conditioning controller needs a feedback mechanism associated with sleep thermal comfort. A thermal comfort prediction system is therefore needed. However, there is a dearth of research in this area. Recent advancements in wearable sensor technology and deep learning provide innovative solutions for monitoring sleep comfort. In this study, we propose a novel attention-based gated recurrent unit neural network (AGRU) designed for the prediction of sleep thermal comfort. This model leverages data from air environment monitors and wearable sensors capable of detecting environmental variables and physiological signals. We conducted an eighty-day sleep experiment involving twenty subjects exposed to varying environmental conditions. The monitors and sensors recorded seven features, including air temperature, relative humidity, skin temperatures at four points, and pulse rate. The results obtained underline the practicality and efficacy of the deep learning model that draws on environmental and physiological signals, with an average accuracy, macro-precision, macro-recall and macro-F1-score of 86.99 %, 87.10 %, 86.98 % and 0.87, respectively. This research provides substantial support for the continued advancement of smart home technology and wearable technology in the field of sleep thermal comfort.

Advancing personal thermal comfort prediction: A data-driven framework integrating environmental and occupant dynamics using machine learning

Various individual thermal preferences do not always lead to optimal thermal comfort, despite the high energy consumption of Heating, Ventilation, and Air Conditioning (HVAC) systems. Thermal comfort models often struggle to accurately predict occupants’ thermal sensations due to their inability to adapt to diverse scenarios. This research employs machine learning techniques to develop personalized data-driven thermal comfort models using a global database of 17,814 samples from the ASHRAE Global Thermal Comfort Database II. Specifically, the study implements random forest and k-nearest neighbor algorithms to analyze the influence of environmental factors (such as air temperature, outdoor temperature, relative humidity, and air velocity) and personal factors (including clothing insulation, activity level, and individual differences) on thermal comfort perception. The study systematically compares the predictive performance of different output types (thermal sensation vs. preference vote), input parameters (6 vs. 12), and machine learning methods. The optimized random forest model with 12 inputs demonstrated an accuracy of over 70% in predicting thermal preference votes, which was significantly higher than the 34% accuracy of the PMV model. The results present new optimized data-driven models and offer insights into the relative influence of parameters such as climate, building cooling strategies, and occupant age and gender. This research aims to enable personalized thermal comfort predictions to improve satisfaction and energy efficiency in various buildings.

Thermal comfort prediction in multi-occupant spaces based on facial temperature and human attributes identification

Research on thermal comfort prediction has garnered significant attention, with numerous studies focusing on predicting individual thermal comfort through various physiological indicators. However, these methods often lack reliability in long-distance, multi-occupant spaces and fail to utilize RGB image information fully. To address these limitations, this study introduces a new method for predicting thermal comfort in multi-occupant spaces by integrating individual attributes and facial temperature features. The research extracts human body attributes from RGB images and temperature features from different facial regions using image registration and deep learning algorithms. Results show that specific facial regions (cheeks, nose, eyes) combined with human attributes effectively predict thermal comfort. In contrast, the forehead, mouth, and chin are less reliable due to their sensitivity to external influences. Among various models, the LightGBM model achieves the highest accuracy at 90.04 ± 4.46%, demonstrating the effectiveness of the model this study proposed. This study offers a novel perspective for thermal comfort prediction in various settings, such as classrooms and lecture halls.

Data-efficient comfort modeling: Active transfer learning for predicting personal thermal comfort using limited data

Personal comfort models are used to predict thermal comfort responses at the individual level rather than predicting the average thermal comfort responses for large populations. These models, data-driven in nature, need to be trained on large amounts of occupant comfort feedback and sensor data to achieve accurate predictions. However, collecting such data is often expensive and labor-intensive in reality. To address this, we proposed a data-efficient active transfer learning (ATL) framework to improve the performance of personal comfort models under limited data. To demonstrate the validity of this framework, we developed a base Convolutional Neural Network-Long Short-term Memory (CNNLSTM) model alongside two transfer learning models utilizing feature extraction (TL-CNNLSTM-FE) and fine-tuning (TL-CNNLSTM-FE) approaches, enhanced by a novel active learning strategy. Using these models, three comfort prediction tasks (i.e., thermal preference, thermal acceptability, and air movement preference) were performed by transferring the knowledge from the ASHRAE Global Thermal Comfort Database II to a limited dataset collected in the tropics. Empirical results indicate that the active transfer learning framework proposed was able to consistently outperform the base and transfer learning models using only less than 10% of the training data for all personal comfort tasks, highlighting the effectiveness of this strategy. The implications of this work are especially useful for the research community working on the practical applications of data-efficient machine learning approaches for personal thermal comfort predictions.

Thermal and visual comforts of occupants for a naturally ventilated educational building in low-income economies: A machine learning approach

Thermal and visual comfort (TVC) significantly impact the health and productivity of occupants as well as the indoor environmental quality in educational buildings. Machine learning approaches have currently gained popularity for TVC prediction; however, accurately predicting TVC faces challenges due to instrumental errors, environmental noise, and imbalanced data. Moreover, studies have revealed the unreliability of TVC predictions, particularly in naturally ventilated buildings. Therefore, this study develops three machine learning models, random forest (RF), decision tree (DT), and XGBoost, to accurately predict TVC using 1310 sets of field survey data collected from a naturally ventilated educational building (NVEB) in Dhaka, Bangladesh. Recursive feature elimination with cross-validation is employed to identify optimal features, while two resampling methods are applied to manage imbalanced classification. The RF model achieves the highest accuracy of 96 % for thermal comfort prediction, while the XGBoost model attains the highest accuracy of 95 % for visual comfort prediction. Meanwhile, the DT model achieves an accuracy of 94 % and 93 % for TVC prediction, respectively. Moreover, both RF and XGBoost outperform DT in class prediction performance, with an average area under the curve (AUC) scores above 0.95, while DT achieves an average AUC score above 0.80. In terms of two resampling methods, the results show that the models developed using the synthetic minority over-sampling technique (SMOTE) + edited nearest neighbour approach leads to less misclassification compared to the SMOTE + Tomek link. These findings demonstrate the capability of machine learning models in accurately predicting TVC in NVEBs and handling imbalanced datasets.

Enhanced Thermal Comfort Prediction Model by Addressing Outliers and Data Imbalance

The relationship between individuals and their thermal environment is pivotal not only for comfort but also for health and productivity. Thermal comfort, as defined by ASHRAE, reflects an individual's satisfaction with their ambient thermal conditions and can be gauged using the ASHRAE scale. In the past, traditional thermal comfort prediction models such as the Predicted Mean Vote (PMV) were used to evaluate thermal comfort. Nevertheless, the emergence of machine learning provides a more dynamic approach to predict thermal comfort of occupants. However, the subjective nature of thermal comfort introduces data ambiguities challenge which lead to the existence of outliers. Moreover, data imbalances within the dataset can cause the machine learning models to not learn the minority class effectively, resulting in the deterioration of the model. This research has developed an enhanced thermal comfort prediction model to predict the occupant’s thermal comfort by leveraging the outlier detection technique and synthetic data generator, particularly the Isolation Forest and SMOTE. The experiment showed that the proposed model is able to achieve an accuracy of 74.94%. This exhibited a slight improvement compared to the findings in prior research of using Random Forest prediction model.

Machine learning thermal comfort prediction models based on occupant demographic characteristics

This study aims to investigate the predictive occupant demographic characteristics of thermal sensation (TS) and thermal satisfaction (TSa) as well as to find the most effective machine learning (ML) algorithms for predicting TS and TSa. To achieve this, a survey campaign was carried out in three mixed-mode buildings to develop TS and TSa prediction models by using six ML algorithms (Logistic Regression, Naïve Bayes, Decision Tree (DT), Random Forest (RF), K-Nearest Neighborhood (KNN) and Support Vector Machine). The prediction models were developed based on six demographic characteristics (gender, age, thermal history, education level, income, occupation). The results show that gender, age, and thermal history are significant predictors of both TS and TSa. Education level, income, and occupation were not significant predictors of TS, but were significant predictors of TSa. The study also found that RF and KNN are the most effective ML algorithms for predicting TS, while DT and RF are the most effective ML algorithms for predicting TSa. The study found that the accuracy of TS prediction models ranges from 83% to 99%, with neutral being the most correctly classified scale. The accuracy of TSa prediction models ranges from 84% to 97%, with dissatisfaction being the most common misclassification.

Modeling indoor thermal comfort in buildings using digital twin and machine learning

Digital Twin (DT) concept is used in different domains and industries, including the building industry, as it has physical and digital assets with the help of Building Information Modeling (BIM). Technologies and methodologies constantly enrich the building industry because the amount of data generated during different building stages is considerable and has a tremendous effect on the lifecycle of a building. Previous research underscores the importance of seamlessly exchanging information between physical and digital assets within a comprehensive framework, particularly emphasizing the integration of BIM data with various systems to enhance efficiency and prevent information loss. Despite advancements in technologies, challenges persist in optimizing methods for integrating BIM data into DT frameworks, including ensuring interoperability, scalability, and real-time monitor and control. This study addresses this research gap by proposing a comprehensive platform that integrates the DT concept with IoT and BIM technologies. The platform is developed in five main stages: 1) acquiring electronic data of the building from the laser scanner, 2) developing a Wi-Fi IoT module and BIM data for physical assets and digital replica, 3) constructing the DT elements of the platform, 4) performing data analysis 5) implementing thermal comfort prediction models. Two machine learning models (Facebook prophet, NeuralProphet) are implemented to predict thermal comfort. The best predictive model is identified by evaluating its error function using historical training data collected during facility operation. A case study demonstrates the practical application of the proposed framework. The case study involves a real building where the platform is implemented to monitor and control indoor environments. By utilizing predefined data in BIM models, the platform ensures data accuracy, consistency, and usability. The case outputs reveal that Neuralprophet provides good prediction results.

Optimising Building Energy and Comfort Predictions with Intelligent Computational Model

Building performance prediction is a significant area of research, due to its potential to enhance the efficiency of building energy management systems. Its importance is particularly evident when such predictions are validated against field data. This paper presents an intelligent computational model combining Monte Carlo analysis, Energy Plus, and an artificial neural network (ANN) to refine energy consumption and thermal comfort predictions. This model addresses various combinations of architectural building design parameters and their distributions, effectively managing the complex non-linear relationships between the response variables and predictors. The model’s strength is demonstrated through its alignment with R2 values exceeding 0.97 for both thermal discomfort hours and energy consumption during the training and testing phases. Validation with field investigation data further confirms its accuracy, demonstrating average relative errors below 2.0% for total energy consumption and below 1.0% for average thermal discomfort hours. In particular, an average underestimation of −12.5% in performance discrepancies is observed when comparing the building energy simulation model with field data, while the intelligent computational model presented a smaller overestimation error (of +8.65%) when validated against the field data. This discrepancy highlights the model’s potential and reliability for the simulation of real-world building performance metrics, marking it as a valuable tool for practitioners and researchers in the field of building sustainability.

The Personalized Thermal Comfort Prediction Using an MH-LSTM Neural Network Method

The Personalized Thermal Comfort Prediction Using an MH-LSTM Neural Network Method

Assessment and prediction of pedestrian thermal comfort through machine learning modelling in tropical urban climate of Nagpur City

Assessment and prediction of pedestrian thermal comfort through machine learning modelling in tropical urban climate of Nagpur City

Thermal comfort study of urban waterfront spaces in cold regions: Waterfront skyline control based on thermal comfort objectives

In waterfronts, thermal comfort is affected by climate and urban morphology. Existing studies on waterfront thermal comfort have mainly focused on warm and hot regions, focusing on the cooling effects of water and vegetation during the summer. The research methods are mostly questionnaire surveys and simulations, whereas the urban form and climate characteristics of cold regions are rarely considered, and the results of these studies are seldom applied to design practices. This study focused on the thermal comfort of waterfronts in cold regions. Computational Fluid Dynamics, Mean Radiant Temperature, and the Universal Thermal Climatic Index were used to examine the effects of tree–water features (winter river ice and tree defoliation) on thermal comfort in cold regions, and correlation analyses were combined to screen for relevant urban morphology factors for waterfront thermal comfort. Regression analyses were conducted to understand the influence of six factors, namely, building orientation, floor area ratio, open space ratio, building height to distance between river and building ratio, standard deviation of the first building row, and vegetation concentration, on waterfront thermal comfort. After training and comparing the four thermal comfort prediction models, a genetic algorithm combined with an artificial neural network was applied to optimize the design of the urban morphology of the six waterfront blocks. The optimization improved the waterfront thermal comfort performance by 22%, 7%, 77%, 106%, 3%, and 3%, respectively.

Prediction of occupant thermal state via infrared thermography and explainable AI

Accurate and real-time assessment of occupant thermal comfort can provide a solid foundation for efficient air conditioning operations. Existing studies already show the feasibility of using contactless technologies for thermal comfort prediction assisted by machine learning algorithms. However, the lack of transparency in machine learning often weakens user trust. This study performs explainable AI analysis to explore the potential of infrared imaging in thermal comfort evaluation. Specifically, an investigation was carried out in a climatic chamber, and infrared cameras were used to collect facial temperature data. Five popular ensemble tree models were employed to construct prediction models, and explainable AI analysis was performed using SHAP (SHapley Additive exPlanations) theory. Results show that combining additional facial information can significantly improve the overall model performance, and certain facial attributes present high contributions based on SHAP values. Combining facial features with explainable AI provides a convincing basis for thermal comfort assessment. The high SHAP values of facial features can also contribute to finding selective occupants with low neutral voting rates, providing evidence for customized cooling or heating from building systems.

Enhancing thermal comfort prediction in high-speed trains through machine learning and physiological signals integration

Heating, Ventilation, and Air Conditioning (HVAC) systems in high-speed trains (HST) are responsible for consuming approximately 70% of non-operational energy sources, yet they frequently fail to ensure provide adequate thermal comfort for the majority of passengers. Recent advancements in portable wearable sensors have opened up new possibilities for real-time detection of occupant thermal comfort status and timely feedback to the HVAC system. However, since occupant thermal comfort is subjective and cannot be directly measured, it is generally inferred from thermal environment parameters or physiological signals of occupants within the HST compartment. This paper presents a field test conducted to assess the thermal comfort of occupants within HST compartments. Leveraging physiological signals, including skin temperature, galvanic skin reaction, heart rate, and ambient temperature, we propose a Predicted Thermal Comfort (PTC) model for HST cabin occupants and establish an intelligent regulation model for the HVAC system. Nine input factors, comprising physiological signals, individual physiological characteristics, compartment seating, and ambient temperature, were formulated for the PTS model. In order to obtain an efficient and accurate PTC prediction model for HST cabin occupants, we compared the accuracy of different subsets of features trained by Machine Learning (ML) models of Random Forest, Decision Tree, Vector Machine and K-neighbourhood. We divided all the predicted feature values into four subsets, and did hyperparameter optimisation for each ML model. The HST compartment occupant PTC prediction model trained by Random Forest model obtained 90.4% Accuracy (F1 macro = 0.889). Subsequent sensitivity analyses of the best predictive models were then performed using SHapley Additive explanation (SHAP) and data-based sensitivity analysis (DSA) methods. The development of a more accurate and operationally efficient thermal comfort prediction model for HST occupants allows for precise and detailed feedback to the HVAC system. Consequently, the HVAC system can make the most appropriate and effective air supply adjustments, leading to improved satisfaction rates for HST occupant thermal comfort and the avoidance of energy wastage caused by inaccurate and untimely predictive feedback.